diff --git a/core/kamon-core/src/main/scala/kamon/trace/Tracer.scala b/core/kamon-core/src/main/scala/kamon/trace/Tracer.scala
index 0eedde6fe..2e8f1ec72 100644
--- a/core/kamon-core/src/main/scala/kamon/trace/Tracer.scala
+++ b/core/kamon-core/src/main/scala/kamon/trace/Tracer.scala
@@ -436,6 +436,12 @@ class Tracer(initialConfig: Config, clock: Clock, contextStorage: ContextStorage
         latencyThresholdNanos = traceConfig.getDuration("tail-sampler.latency-threshold").toNanos
       )
 
+      if(tailSamplerSettings.enabled && delayedSpanReportingDelay.isZero) {
+        _logger.warn(
+          "Enabling tail sampling without any sensible span-reporting-delay setting will probably lead to incomplete " +
+          "traces. Consider setting span-reporting-delay to a value slightly above your application's timeout setting")
+      }
+
       if(_traceReporterQueueSize != traceReporterQueueSize) {
         // By simply changing the buffer we might be dropping Spans that have not been collected yet by the reporters.
         // Since reconfigures are very unlikely to happen beyond application startup this might not be a problem.